The widespread use of rotating electrical machines makes it necessary to get a true and exact picture of their condition. Various methods exist for making it possibele to detect failures in the machines. Either when measuring electrical parameters, or analysing the vibration, (one important part of this dissertation), in all cases it is essential that the person, making the diagnosis is aware of the well working machine's characteristics.
Based on the development process of each failures, causes and causations can be observed, which may provide the basis for more serious problems. For example, the damaged shaft may lead to failure of the bearings, or the mechanical vibrations caused by looseness can cause damage to other machine elements as well. Because there are always some unwanted forces acting when failures occur, vibrations will be generated with varying intensity, depending on the stiffness of the structural elements.
By measuring the vibrations of the machine we can deduce the machine-condition, as well as the individual components' status. Most of the errors (rotor imbalance, eccentricity, misalignment, damaged bearings, etc) are clearly identifiable. By separating the noticeable harmonic vibration components the various vibrations of the machine parts are separable, so that the vibration generating elements can be determined. Years of experience are necessary for being able to set up a good diagnosis after the evaluation of the signals.
The purpose of this study is to describe the relationship between the faults in rotating electrical machines and to present a diagnosis method based on vibration measurements.